Working Class Votes and Conservative Losses: Solving the UKIP Puzzle

Transcription

1 Parliamentary Affairs Advance Access published April 17, 2015 Parliamentary Affairs (2015) 1 16 doi: /pa/gsv005 Working Class Votes and Conservative Losses: Solving the UKIP Puzzle Geoffrey Evans * and Jon Mellon Department of Politics, Nuffield College, University of Oxford, Oxford OX1 1NF, UK * Correspondence: geoffrey.evans@nuffield.ox.ac.uk Opinions are divided on whether the Conservatives or Labour need to worry most about UK Independence Party (UKIP) in the 2015 General Election. How do we reconcile evidence of substantial levels of UKIP support among traditional working class voters, and in Labour constituencies, with evidence that UKIP voters report voting Conservative in 2010? In this article, we resolve this implicit contradiction using long-term panel data to examine the sequencing of vote switching from Labour to UKIP. We argue that Labour s move to the liberal consensus on the EU and immigration led to many of their core voters defecting before UKIP were an effective political presence. We show that not only is the working-class basis of UKIP overstated but the party is mainly attracting disaffected former Labour voters from the Conservatives and elsewhere, which is why the Conservatives, not Labour, will feel most of the electoral pain in Keywords: UKIP (UK Independence Party), Voting, Working class, British politics, UK Conservative Party 1. Introduction Numerous journalistic reports have pointed to evidence that the Conservative Party appear to be most threatened by The UK Independence Party (UKIP). Opinion polls consistently find that it is that people who report voting Conservative in 2010 who are most likely to have shifted to UKIP. To give just some examples: analysis of numerous opinion polls by John Curtice in May 2013 showed that 16 per cent of those who voted Conservative in 2010 and currently have a party preference would now vote UKIP. The equivalent figure for the Liberal Democrats is 8 per cent and for Labour 4 per cent (Curtice, 2013). Peter Kellner s (2014) analysis of local election outcomes found that if voting followed a similar pattern in a general election UKIP would have won nine parliamentary seats all currently held # The Author Published by Oxford University Press on behalf of the Hansard Society; all rights reserved. For permissions, please journals.permissions@oup.com

2 Page 2 of 16 Parliamentary Affairs by the Conservatives. In 2014 Lord Ashcroft found that 52% of UKIP voters in the European elections reported voting Conservative in 2010, with only 15% reporting previously voting Labour (Ashcroft Polls, 2014). Likewise a report using the British Election Panel Study found that 43.5% of UKIP voters had voted Conservative in 2010, and only a mere 12.9% Labour (Evans and Mellon, 2014). These figures and many other examples indicate conclusively that UKIP draws vastly more support from Conservative voters than from those of any other party. This electoral support is accompanied by a similar sentiment among Conservative Party members (Webb and Bale, 2014) and by accompanying strategic responses by the party itself (Lynch et al., 2012). Moreover this recent pattern should not surprise us too greatly: the Conservative Party has a history of internal division and vote loss on the EU going back to the mid-1990s when the Party was last in power prior to 2010 (Evans, 1998) and the impact of James Goldsmith s Referendum Party on Conservative support in the 1997 election (Heath et al., 1998). The EU has been a constant thorn in the unity of the party. At the same time, it has been commonly observed that working-class voters are attracted to so-called radical right-wing (RRP) parties such as UKIP in the many European countries in which RRPs have flourished. It should not surprise us therefore that some recent observers of British electoral behaviour have emphasised the significance of the working class for UKIP s current popularity and likely vote attainment in the forthcoming election. They have also argued that this poses a serious threat to the Labour Party rather than the Conservatives. It is assumed that a party with a strong basis of support in the working class will hurt, electorally, the main party (Labour) that traditionally represents the working class in Westminster. In this article we show that this is not the case. Both notions: that UKIP appeal to those on the right and are thus a threat to the Conservatives, and at the same time draw upon substantial support from the working class, are to some degree true. What we show not to be true is the assumption that most workingclass defection to UKIP is working-class defection from Labour. In part this is because what some commentators have labelled the working class include the hard-core, right-wing, self-employed, who have traditionally been even more Conservative than the professional and managerial middle classes. Additionally, however, we show that the drift away from Labour by its traditional workingclass supporters occurred before the current electoral cycle and is unlikely to be easily reversed given the party s recent shifts in policy and ideological stances (Evans and Tilley, 2012), and social composition (Heath, 2015). Instead of the Conservatives, the Liberal Democrats, or not voting, former Labour voters are now attracted to UKIP. To a fair degree, therefore, UKIP s increased presence in the competition for votes in 2015 just reshuffles the opposition vote in Labour seats.

3 Solving the UKIP Puzzle Page 3 of UKIP as a working-class party In Revolt on the Right: Explaining Public Support for the Radical Right in Britain Ford and Goodwin (2014a) present an important, and at first glance given the consensus on the Conservative basis of UKIP support noted above counter-intuitive analysis. Rather than being a threat to the Conservative Party (e.g. Hayton, 2010; Whitaker and Lynch, 2011), UKIP draws much of its support from working-class voters alienated by a collapse in living standards and the lack of good jobs. Ford and Goodwin argue that UKIP attracts working-class, white, male voters because they are the losers in Britain s rapid economic and social transformation and the centrist politics advocated by the Westminster liberal elite: in Ford and Goodwin s words they are the left behind ; older, working class voters with few qualifications. For them, UKIP is a working-class phenomenon. Its support is heavily concentrated among older, blue-collar workers, with little education and few skills; groups who have been left behind by the economic and social transformation of Britain in recent decades, and pushed to the margins as the main parties have converged in the centre ground. UKIP are not a second home for disgruntled Tories in the shires; they are a first home for angry and disaffected working class Britons of all political backgrounds, who have lost faith in a political system that ceased to represent them long ago. (Ford and Goodwin, 2014a, p. 270) The theme of the shrinking size and consequent political marginalisation of the British working class is not new (see Evans et al., 1999; Tilley and Evans, 2011; Evans and Tilley, 2012), nor is the thesis that the EU has provided a focal point for class realignment (Evans, 1999), but the application of the logic of this process to UKIP voting and the collation of supporting empirical evidence is an important contribution. From a broader European perspective, moreover, this theme is well established. Numerous comparative studies have shown high support for RRPs among the working class (Betz, 1994; Kitschelt, 1995; Houtman, 2003; Ivarsflaten, 2005; de Lange, 2007; amongst others). The key issues that characterise such voters are Euroscepticism, opposition to immigration, support for punitive law-and-order policies and general political disaffection. In Britain such preferences have long been known to separate the working class from the middle class Euroscepticism in particular being noticeably more divisive than even traditional economic preferences on redistribution (Evans, 2000). It would be surprising therefore if this thesis did not also apply to Britain. Comparative research into class realignment also finds that these voter shifts can be motivated by the changing programmatic appeals of left-wing parties who move to centrist positions (Kitschelt and McGann, 2005; Achterberg and

4 Page 4 of 16 Parliamentary Affairs Houtman, 2006; Rennwald and Evans, 2014). Most comprehensively Spies (2013) examines support for RRPs among working-class voters in 13 West European societies for the periodfrom 1980 to He finds that in countries where theeconomic dimension of party competition has decreased in both salience and polarisation, the support for RRPs among the working class is considerably higher than in countries that do not show such a trend. In the British case, the realigning impact of Labour s shift to a pro-eu position (following its 1989 policy review) and the consequent weakening of working-class commitment to Labour was established in the 1990s (Evans, 1999). The conditions for cross-class voting have been in place for some time and the emergence of UKIP fits well with this process. However, there is a crucial weakness in drawing inferences from these well-established processes for the impact of UKIP voting in the 2015 election: the sequencing of voter transitions. 2.1 Our argument The impact of sequencing Our primary argument is that Labour s move to the liberal consensus on the EU and immigration alienated many of their core voters a long time before UKIP were an effective political presence. These disaffected core voters left Labour in 1997, 2001, 2005 and 2010 and went to other parties or simply stopped voting. UKIP has since attracted these disaffected former Labour voters, particularly from the Conservatives. As a result of this pattern of recruitment, UKIP will not be a great threat to Labour seats in 2015; they will steal more votes from the non-labour parties in those seats (or people who did not vote at all in 2010). Hence the mismatch between constituency-level findings, which have shown substantial UKIP support in Labour-held seats, and individual voters reported patterns of defection which have mainly been from the Conservatives. In brief, the damage to Labour s core support had already been done by new Labour s focus on a pro-middle class, pro-eu and, as it eventually turned out, pro-immigration agenda, before the arrival of UKIP as a plausible electoral choice in the years following the 2010 election. Prior to then, Eurosceptic voters, whether working class or otherwise, could switch to the Conservatives as they were not in power and could more easily present themselves as an effective Eurosceptic party relative to Labour or the Liberal Democrats. Consistent with this argument, Evans (2002) analysis of the electoral cycle found that EU-scepticism was the only significant predictor of vote switching from Labour to the Conservatives in the 2001 election. Further analyses of the electoral cycle similarly found that the Government s handling of immigration was an important predictor of defection from Labour in the 2010 election (Evans and Chzhen, 2013). Once in power, however, the Conservatives were to some degree implicated in the EU project. Most obviously their hands were to a large degree tied on immigration by EU legislation and the

5 Solving the UKIP Puzzle Page 5 of 16 free movement principle. This will have opened space for UKIP after 2010 to be the only party still able to realistically reject EU membership and its implications for immigration from the EU. In contrast to the votes lost by Labour, the electoral costs of the EU/immigration for the Conservatives are occurring in the current electoral cycle Misrepresenting the class basis of UKIP support Our second point is that the significance for Labour of what might be called blue collar switching to UKIP has been exaggerated by Ford and Goodwin s glossing over important distinctions in the types of class basis of UKIP voters. There are two quite distinct social groups that have shown a disproportionately high level of support for RRP parties: the working class and the somewhat quaintly labelled petty bourgeoisie (the self-employed small employers such as shop owners). Empirical evidence of RRP support among both of these classes has been demonstrated in many studies (Kitschelt, 1995; Lubbers et al., 2002; Ignazi, 2003; Ivarsflaten, 2005; Kitschelt and McGann, 2005; Arzheimer and Carter, 2006; Rydgren, 2012). Ivarsflaten (2005) incisively shows that working-class and petty-bourgeoisie RRP voters are divided on economic issues, but share the types of non-economic preferences addressed by RRPs. Unfortunately, even Ford and Goodwin s (2014a, pp ) presentation of data on class and UKIP voting using the Goldthorpe Heath class schema: the most robust and theoretically well-grounded way to measure social class (Ford and Goodwin, 2014a, p. 292) does not include the petty bourgeoisie. This may be because the primary basis of Ford and Goodwin s analysis of support for UKIP, YouGov s Continuous Monitoring Survey, uses the Social Grade classification, which does not allow the identification of the selfemployed and small business owners. In fact, 21% of respondents in the British Election Study (BES) who are classified as skilled manual workers in Social Grade are actually self-employed, the highest proportion among any of the Social Grade categories. In any study of RRP support, however, this is an unfortunate omission: the key fact of interest about the self-employed and small business employers is not just their blue collar-ness (as many are) but the intensity of their support for the Conservatives, which goes back as far as 1964 when the first British Election Study survey was undertaken (Heath et al., 1991). Any shift by this group towards UKIP overwhelmingly hits the Conservatives, not Labour. 3. Empirical analysis We start by examining the sequencing of transitions from Labour to UKIP. To demonstrate the importance of sequencing, Figure 1 shows the flow of the vote to UKIP from 2005 to 2014 using respondents who were in the 2005 and 2010 BES panel

6 Page 6 of 16 Parliamentary Affairs Figure 1 Pathways to UKIP surveys 1 as well as those in the third wave (autumn 2014) of the current BES panel study. 2 Note that we only show respondents who end up supporting UKIP in 2014 and exclude all other respondents from the diagram. We can see that UKIP s voters are overwhelmingly taken from those who voted Conservative in Even the Liberals lose more to UKIP than do Labour. However, we can also see that 2005 Labour voters also gravitated to UKIP. What distinguishes them is that they did so via more circuitous routes. Figure 2 shows that a clear majority of 2005 Labour voters who supported UKIP in 2014 had left the party by To summarise, 80% of 2005 Conservatives who intended to vote for UKIP in 2014 were still Conservative in 2010, whereas a mere 40% of 2005 Labour voters who supported UKIP in 2014 were still Labour. So Labour s loss to UKIP is partly hidden by its supporters having gone via other paths prior to arriving at UKIP. Moreover, these defecting Labour voters are to a large degree the sorts of 1 Note that all theanalysis here focuses on England and Wales, due to the potentiallycomplicating effect of the rise of the Scottish National Party over this same time period. 2 These figures use recalled profile data for the 2010 and 2005 vote. The 2010 vote data were primarily collected in 2011 and the 2005 vote data were mainly collected in 2009, with some data being collected later in both cases. The 2010 vote data have 94% agreement with the recorded vote choice for those panel respondents who took the British Election Study post survey in The equivalent figure for 2005 is 79%.

7 Solving the UKIP Puzzle Page 7 of 16 Figure 2 Percentage of UKIP intenders who supported the same party in 2005 and 2010 by 2005 support. people that new Labour will have alienated. Figure 3 shows predictable patterns: the working class, 3 employers and anti-eu 2005 Labour voters tend to be those who ended up supporting UKIP. That the EU and immigration eventually merged as an issue, particularly after the 2004 accession of former Communist countries, is also indicated by the similarity in the patterns of attitudes towards immigration and Europe shared by those who did or did not switch from Labour to UKIP. We next examined these inferences using multivariate models of vote (shown in Table 1). We run a multinomial logistic regression predicting 2014 vote intention among 2010 Conservative supporters. 4 These models show that 2010 Conservatives who were 2005 Labour voters are more likely than other Conservatives to switch to UKIP. Routine workers are also more likely to switch to UKIP. Moreover, 3 We combine classes 5 7 from the National Statistics Socio-economic classification (NS-SEC) analytic classes. These very closely correspond to the equivalent Goldthorpe Heath classes. 4 The models shown here use unweighted data, following Gelman s (2007) discussion that weighting can reduce the efficiency of estimation. All the regression results in this article are robust to using survey weights.

8 Page 8 of 16 Parliamentary Affairs Figure 3 Sources of defection to UKIP from Labour: Class, Anti-EU and anti-immigration attitudes (2010) among 2005 Labour supporters broken down by their 2014 party preference. Table 1 Multinomial logit model predicting 2014 vote intention among 2010 conservative voters UKIP Labour Coef Std error p Coef Std error p 2005 vote Non-voting Labour Lib Dem UKIP Other Labour handle immigration (2010) Conservatives handle immigration (2010) Constant Base category: Conservative. An other outcome category was also included (not shown). n ¼ 1485, log likelihood ¼ whereas Conservative/Labour switching is unrelated to either party s handling of immigration, 5 Conservative/UKIP switching is strongly (negatively) 5 Handling of immigration is measured using the questions How well do you think the present government has handled immigration? and How well do you think a Conservative government would handle immigration? Both questions were asked in the 2010 pre-campaign British Election Study online panel survey. Respondents answer on a 5-point Likert scale from Very well to Very badly. These questions allow us to track perceptions of the major parties handling of immigration prior to UKIP s major gains in vote intention.

9 Solving the UKIP Puzzle Page 9 of 16 Figure 4 Percentage support for UKIP among different social classes (excluding respondents who stated do not know for vote intention). predicted by 2010 perceptions of both Conservative and Labour handling of immigration. These analyses add support to the argument that those supporters who left Labour over their handling of immigration moved onto UKIP and that Labour defectors who initially went to the Conservatives are prime targets for UKIP. Thus far we have seen how the sequencing of voters defection from Labour has influenced the impact of UKIP support in such a way as to hurt the Conservatives rather than Labour in the General Election. A further reason why the emphasis on UKIP being a problem for Labour is misplaced is the failure to differentiate the hard core, right-wing self-employed from the working class. Figure 4 presents a more detailed breakdown of the class basis of UKIP support using NS-SEC categories

10 Page 10 of 16 Parliamentary Affairs Table 2 Logistic regression predictors of voting for UKIP (base category all others. Base category for class is low professional) Coef Std error p Coef Std error P Owners Managers High professional Low professional (reference) Self-employed Intermediate Lower supervisory Semi-routine workers Routine workers EU approval (2010) Anger about immigration (2010) Constant Log likelihood n which are derived from the Goldthorpe Heath schema (Rose and Pevalin, 2003) and are thus very similar to those advocated by Ford and Goodwin. 6 Figure 4 shows that employers (class L1), self-employed (4) and lower supervisory (5), semi-routine workers (6) and routine workers (7) have the highest levels of support for UKIP. The idea that UKIP are picking up primarily left behind old Labour supporters is not supported by this more detailed evidence on the class basis of UKIP support. There is strong support amongst those in routine occupations, but some of their strongest support comes from large employers and the selfemployed, who were Mrs Thatcher s hard-core supporters, not Labour s. We find for example that 62% of UKIP supporters think that Thatcher was good for Britain, compared with 44% in the population as a whole, 13% of Labour voters and 87% of Conservative voters. Their perceptions of Thatcher are closer to those of the Conservatives than to Labour. Moreover, the lower supervisory category the strongest UKIP supporters are clearly not the semi- and unskilled manual workers that characterise the left behind working-class UKIP supporters in Ford and Goodwin s thesis. This point is reiterated and strengthened when we look at the drivers behind the UKIP vote. Support for UKIP has been thought to have its roots in anger about immigration and, of course, the EUitself, as a source of threat to jobs, local services and cultural traditions. These are widespread concerns and provide an intelligible way of interpreting UKIP voting. We therefore included measures of anger about 6 We use the NS-SEC analytic categories, but break out large employers and higher managers separately.

11 Solving the UKIP Puzzle Page 11 of 16 immigration as well as approval of EU membership (both measured in 2010) in logistic regression models predicting UKIP voting versus voting for other parties. These models are shown in Table 2, which compares a class-only prediction of UKIP support with a model that includes class and EU/immigration attitudes. The first model shows the same pattern of UKIPsupport as we saw in the descriptive statistics, with owners, self-employed and working-class voters having significantly higher levels of UKIP support. When responses to the attitudinal questions are included in our models, European preferences and anger about immigration are, unsurprisingly, highly statistically significant. However, the inclusion of these controls eliminates the positive relationship between the working class and UKIP intention. This suggests that working-class support for UKIP may be primarily driven by anti-immigration and anti-eu attitudes. In contrast, even after controlling for 2010 EU and immigration concerns, the self-employed are still significantly more likely to vote UKIP. 7 The self-employed would appear to be somewhat more solidly UKIP than the left behind working class themselves. Their higher levels of support for UKIP are more deeply rooted than the concerns about immigration and the EU that can explain working-class support for UKIP. 8 This certainly does not look like a simple case of the white working class supporting UKIP. Much of their support is from the Conservative heartlands. 9 Overall, the picture is nearer to Ivarsflaten s (2005) description of an alliance between the working class and the self-employed (and employers) than to a party of the disadvantaged. Finally, if we look at where UKIP s support comes from rather than from differences between classes in levels of support we see an arguably even more interesting story. While UKIP disproportionately appeals to working-class groups and self-employed or small employer classes, Figure 5 shows that they receive a greater proportion of their support from lower professionals and managers (class 2) than from any other class. 10 This of course reflects the fact that this group is the largest NS-SEC analytic class in the UK. As with most other political parties in Britain, UKIP draws support from across many different classes. Again, the notion of a party of the left behind is not supported. 7 Significance is tested at the 10% level here. The sample sizes for this comparison are relatively small (4636) because we only look at respondents who took part in the 2010 and 2015 British Election Study panel. 8 We get substantively similar results using multinomial logit models of vote, using different reference categories for class and using survey weights in the regression. 9 In further multivariate analyses not shown here we demonstrate that private sector employees also have high levels of UKIP support. 10 Note that we reweight the British Election Study data to national NS-SEC analytic class census distributions for this figure.

12 Page 12 of 16 Parliamentary Affairs Figure 5 Percentage 2014 UKIP supporters and other supporters coming from each social class (excluding respondents who stated do not know for vote intention). 4. Conclusions Our argument is that Labour s move to the liberal consensus on the EU and immigration left many of their core voters out in the cold a long time before UKIP were an effective political presence. To some degree therefore what UKIP are doing is picking up already disaffected former Labour voters from the Conservatives and elsewhere: for example, a quarter of those who left the Conservatives for UKIP after 2010 had only voted Conservative in that year, and the largest group of these had previously voted Labour in The implication is that UKIP will not be a great threat to Labour s seats: they will obtain more votes from the opposition parties in those seats, or from people who had not voted in Hence the mismatch between recent constituency level findings of fairly high levels of support for UKIP in Labour seats, despite individual voters having defected primarily from the Conservatives. So if we use the BES panel study to examine the sources of UKIP support in Labour seats we see that

13 Solving the UKIP Puzzle Page 13 of 16 (1) Only 20% voted Labour in 2010, whereas 44% were Conservatives. (2) In % of these voters supported Labour and only 38% were Conservative. 11 So the answer to why UKIP are doing well in Labour seats but are unlikely to win these seats is that there are a lot of Conservative voters in them (25% of the vote in 2010 on average), some of whom are Labour defectors. And as shown above, multivariate analysis indicates UKIP have picked up support disproportionately from these former Labour voters. We should note that this possibility was suggested by Ford and Goodwin (2014a, p. 171), who lacked the data to evaluate it, and the implication it holds for the 2015 election is not integrated into their general analysis. Ford and Goodwin claim UKIP is hurting Labour because Tory voters who migrated to UKIP should be returning to Labour at this point in the electoral cycle. But at the same time they also argue that this process is part of a longer-term realignment between the Labour Party and traditional Labour voters who are anti-eu and anti-immigration, and more generally socially conservative: this defection is thus not a short-term response, and as such the assumption that these sorts of voters would have drifted back to Labour in 2015 is implausible. Labour has not changed its demeanour or policy programme in such a way as to appeal to such voters. It is more likely that disillusionment with the Coalition would have led to their not voting a pattern that many workingclass, former Labour voters displayed during the convergence years when Labour nailed itself to the flag of Blairism in an attempt to increase its electoral competitiveness (Tilley and Evans, 2011). Ford and Goodwin also suggest that the main threat to Labour is likely to come in 2020: Let us be clear: Farage and UKIP are unlikely to win seats in Labour territory in 2015, not least because the MPs often enjoy formidable majorities. But through the local, European and general elections UKIP could quite easily establish themselves as the second political force in many of these areas, positioning themselves as potential challengers to an unpopular Labour government in 2020 (Ford and Goodwin, 2014b). In the spirit of such speculation, however, we might just as easily foresee that by then the Conservatives could be in opposition, with a new leader, and less tarnished by the restrictions of government with respect to the EU. In such circumstances UKIP could well lose many of the large proportion of former Conservatives that will probably swell its ranks in Note that these figures only include respondents who voted in 2010 and exclude non-voters/don t knows and ineligible respondents.

14 Page 14 of 16 Parliamentary Affairs In summary, Ford and Goodwin s argument that UKIP is dividing the left more than the right (Ford and Goodwin, 2014c) and replacing Labour as the main party of the working class misses the significance of the sequencing of voter defections: labour drove these people away before UKIP arrived. But we should also note that UKIP s rise has to be understood in the context not only of the Labour Party s move to the centre and its impact on their core support, but also in the Conservative Party s own resulting centrist shift, which will have in turn alienated some of its core supporters. The extent of support for RRPs by right-wing groups such as small business employers and the self-employed has been observed in many other European societies even in 1930s Germany (Hamilton, 1983). These people have been overlooked in Ford and Goodwin s analysis, as has the rather more prosaic observation that most UKIP support actually comes from the established middle classes, if only because these are the largest classes. These are clearly not the left behind. We conclude then that UKIP will hurt the Conservatives in the upcoming General Election. This should not surprise greatly as even Ford and Goodwin s own data show that in , 47% of UKIP support came from the Conservatives and only 7% from Labour (2014a, p. 166). More concretely our own calculations suggest that assuming levels of vote intention remain similar to those measured in late 2014: in the forthcoming General Election the Conservatives could lose around two million of their 2010 voters to UKIP, whereas the Liberal Democrats will lose about 750,000, and Labour only 600,000 (19, 11 and 7%, respectively). The implications of these figures are disturbing for the Conservatives, if we believe that UKIP voters will vote this way in the General Election. Or alternatively, it could beavery promisingonefor them ifwe thinkthatukipvoters willswingback to their longer-term parties in As it is, evidence from the BES panel indicates substantial stability in UKIP support across the three waves conducted in 2014 at both aggregate and individual levels, with UKIP voters expressing almost as much certainty about their vote intention as Conservative intenders. 12 It suggests voters probably do not intend to switch back. In that case the Conservatives will suffer and Labour will benefit with, ironically, the likely outcome being no EU referendum. Funding This article uses evidence collected as part of the 2015 British Election Study funded by the ESRC (Grant number: ES/K005294/1). References Achterberg, P. and Houtman, P. (2006) Why do so Many People Vote Unnaturally? A Cultural Explanation for Voting Behaviour, European Journal of Political Research, 45, Certainty is measured on a 1 7 scale. UKIP voters have a mean certainty score of 6.0 compared to a mean score of 6.2 for Conservative voters in wave 3 of the BES panel.

15 Solving the UKIP Puzzle Page 15 of 16 Arzheimer, K. and Carter, E. (2006) Political Opportunity Structures and Right-Wing Extremist Party Success, European Journal of Political Research, 45, Ashcroft Polls (2014) Post European Election Poll (24 May), wp-content/uploads/2014/05/lord-ashcroft-polls-post-euro-election-poll- Summary-May-2014.pdf. Betz, H.-G.(1994) Radical Right-Wing Populism in Western Europe, NewYork, St. Martin spress. Curtice, J. (2013) If the Conservatives Are to Counter the Challenge of UKIP, They Need to Turn Around the Economy Rather Than Fret Obsessively about Europe (1 May), de Lange, S. L. (2007) A new Winning Formula? The Programmatic Appeal of the Radical Right, Party Politics, 13, Evans, G. (1998) Euroscepticism and Conservative Electoral Support: How an Asset Became a Liability, British Journal of Political Science, 28, Evans, G. (1999) Europe: A New Electoral Cleavage?. In Evans, G. and Norris, P. (eds) Critical Elections: British Parties and Voters in Long-Term Perspective,London,Sage,pp Evans, G. (2000) The Working Class and New Labour: A parting of the ways?. In Jowell, R., Curtice, J., Park, A., Thomson, K., Bromley, C., Jarvis, L. and Stratford, N. (eds) British Social Attitudes, the 17th Report: Focusing on Diversity, London, Sage, pp Evans, G. (2002) European Integration, Party Politics and Voting in the 2001 Election, Journal of Elections, Public Opinion & Parties, 12, Evans, G. and Chzhen, K. (2013) Explaining Voters Defection from Labour Over the Electoral Cycle: Leaders, Economics, and the Rising Importance of Immigration, Political Studies, 61, Evans, G. and Mellon, J. (2014) All Roads Lead to UKIP?, British Election Study Website (9 December), Evans, G. and Tilley, J. (2012) The Depoliticization of Inequality and Redistribution: Explaining the Decline of Class Voting, The Journal of Politics, 74, Evans, G., Heath, A. and Payne, C. (1999) Class: Labour as a Catch-All Party?. In Evans, G. and Norris, P. (eds) Critical Elections: British Parties and Voters in Long-Term Perspective, London, Sage, pp Ford, R. and Goodwin, M. (2014a) Revolt on the Right: Explaining Support for the Radical Right in Britain, London, Routledge. Ford, R. and Goodwin, M. (2014b) Why UKIP and the Radical Right Matter for Progressives, Policy Network, 10 April. Ford, R. and Goodwin, M. (2014c) UKIP Has Divided the Left, not the Right, and Cut Labour off From its Old Support, Guardian, 16 May. Gelman, A. (2007) Struggles with Survey Weighting and Regression Modeling, Statistical Science, 22, Hamilton, R. (1983) Who Voted For Hitler? Princeton, Princeton University Press.

16 Page 16 of 16 Parliamentary Affairs Hayton, R. (2010) Towards the Mainstream? UKIP and the 2009 Elections to the European Parliament, Politics, 30, Heath, O. (2015) Policy Representation, Social Representation and Class Voting in Britain, British Journal of Political Science, 45, Heath, A. F., Jowell, R., Curtice, J., Evans, G., Field, J. and Witherspoon, S. (1991) Understanding Political Change: The British Voter , Oxford, Pergamon. Heath, A., Jowell, R., Taylor, B. and Thomson, K. (1998) Euroscepticism and the Referendum Party, British Elections & Parties Yearbook, 8, Houtman, D. (2003) Lipset and Working-Class Authoritarianism, The American Sociologist, 34, Ignazi, P. (2003) Extreme Right Parties in Western Europe, New York, Oxford University Press. Ivarsflaten, E. (2005) The Vulnerable Populist Right Parties: No Economic Realignment Fuelling Their Electoral Success, European Journal of Political Research, 44, Kellner, P. (2014) How Ukip Will Win the Election, Prospect Magazine, 20 February, Kitschelt, H. (1995) The Radical Right in Western Europe: AComparative Analysis, Ann Arbor, MI, The University of Michigan Press. Kitschelt, H. and McGann, A. J. (2005) The Radical Right in the Alps: Evolution of Support for the Swiss SVP and Austrian FPO, Party Politics, 11, Lubbers, M., Gijsberts, M. and Scheepers, P. (2002) Extreme Right-Wing Voting in Western Europe, European Journal of Political Research, 41, Lynch, P., Whitaker, R. and Loomes, G. (2012) The UK Independence Party: Understanding a Niche Party s Strategy, Candidates and Supporters, Parliamentary Affairs, 65, Rennwald, L. and Evans, G. (2014) When Supply Creates Demand: Social-Democratic Parties Electoral Strategies and the Evolution of Class Voting, West European Politics, 37, Rose, D. and Pevalin, D. (eds) (2003) A Researcher s Guide to the National Statistics Socio- Economic Classification, London, Sage. Rydgren, J. (ed.) (2012) Class Politics and the Radical Right, London, Routledge. Spies, D. (2013) Explaining Working-Class Support for Extreme Right Parties: A Party Competition Approach, Acta Politica, 48, Tilley, J. and Evans, G. (2011) The New Class Divide in Turnout: How Labour s Move to the Centre Has Led to the Political Exclusion of the Less Privileged, Paper presented at the EPOP Annual Conference, Exeter. Webb, P. and Bale, T. (2014) Why Do Tories Defect to UKIP? Conservative Party Members and the Temptations of the Populist Radical Right, Political Studies, 62, Whitaker, R. and Lynch, P. (2011) Explaining Support for the UK Independence Party at the 2009 European Elections, Journal of Elections, Public Opinion and Parties, 21,